judgement and decison
Category and Outcome in Models
Definitions and Concepts
Category: A distinct group or class to which specific instances belong.
Instance: A specific occurrence or example that can be analyzed or classified.
Representativeness: The degree to which an instance resembles the category it belongs to.
Representativeness as a Heuristic
Functionality of Representativeness
According to the concept of representativeness heuristics, likelihood or probability is determined by the extent to which an instance is representative of a category.
If instance 'a' resembles category 'b' closely, the likelihood of 'a' being from 'b' is perceived as high.
Conversely, if 'a' does not resemble 'b', the perceived probability of 'a' originating from 'b' is low.
Application of Heuristics to Examples
Case Study of Bill
Participants are presented with options regarding Bill's likelihood of engaging in activities (e.g., playing jazz or being an accountant).
Participants evaluate the probability based on how closely Bill's description resembles their stereotype of the roles provided.
Case Study of Linda
Similar methodology is used with Linda, where participants estimate her likelihood of being in different categories (e.g., feminist vs. bank teller).
Linda’s description suggests she is more likely to be a feminist due to characteristics such as being outspoken and intelligent, reflecting confirmation bias in casting her as a feminist rather than a neutral bank teller.
Findings and Probability Correlation
Research Outcomes
The correlation between representativeness and subjective probability was reported to be 96% for Bill and 98% for Linda, indicating that judgments based on representativeness frequently surpass logical probability considerations.
This phenomenon is exemplified through 'conjunction fallacy.'
Conjunction Fallacy
Definition
Conjunction Fallacy: The mistaken belief that specific conditions are more probable than a general one.
Explanation using Logical Principles
It is logically impossible for a specific conjunction (e.g., "Linda is a bank teller and an active feminist") to be more likely than the broader general category (e.g., "Linda is a bank teller").
Example Explanation
The intersection of a subset (feminists who are bank tellers) is always less or equal in size to the parent category (employees who are bank tellers). This leads to the logical conclusion that such conjunctions cannot exceed the probabilities of their individual categories.
Additional Research and Misinterpretations
University Study on Probability Perceptions
Research conducted at the University of British Columbia revealed that 85% of participants incorrectly deemed the conjunction of two characteristics as more likely than a singular characteristic.
Further manipulations, including breaking down logical implications, did not sufficiently correct misunderstanding about probabilities in nearly all participants.
General Misunderstandings and Psychological Mechanisms
Insensitivity to Base Rates
Participants often failed to consider base rates when evaluating probabilities, focusing instead on resemblance.
Example of Jack: Despite his attributes fitting more with an engineer due to personal interests, the base rate of lawyers in his workplace could logically suggest he is more likely to be a lawyer.
Misconception of Chance
Participants might perceive certain outcomes as less likely based on representativeness rather than actual statistical probability.
Example: In flipping a coin, patterns perceived as more typical (e.g., alternating heads and tails) can actually mislead assessment of likelihood, as all configurations remain equally probable in random trials.
Ongoing Studies and Their Implications
Recent Study on Various Occupations and Characteristics
Participants were presented with sketches of a character named Richard with varying occupations (teacher, rapist, Christian) and their likelihood estimations.
The study aimed to measure conjunction fallacy occurrences, observing how stereotypes influenced perceived probabilities in relation to professions.
Wrap Up
Final Discussion Points
The introduction of concepts such as conjunction fallacy and insensitivity to base rates fosters a deeper understanding of how people often deviate from logical reasoning in probabilistic contexts.
Evaluating individual examples underscored the challenge in correcting misconceptions rooted in representative heuristics.
Category and Outcome in Models
Definitions and Cognitive Concepts
Category: A distinct mental grouping or class (schema) to which specific instances belong based on shared attributes.
Instance/Exemplar: A specific occurrence, person, or example retrieved from memory that can be analyzed, classified, or compared against a prototype.
Representativeness: The cognitive degree to which a specific instance resembles the "prototype" or the average central tendency of the category it belongs to.
Prototype Theory: Suggests that people categorize items based on how closely they match a mental image of the "typical" member.
Representativeness as a Heuristic
Functionality and Cognitive Efficiency
Developed by Amos Tversky and Daniel Kahneman, the representativeness heuristic is a mental shortcut where probability is judged by the similarity of an object to a category.
Heuristic Mechanism: Instead of performing complex Bayesian calculations, the mind asks: "How much does look like ?"
If instance resembles category closely (high similarity), the subjective likelihood of originating from is perceived as high ().
Conversely, if deviates significantly from the stereotype of , the perceived probability is low, even if statistical evidence suggests otherwise.
Application of Heuristics to Examples
Case Study of Bill (The Stereotypical Accountant)
Participants were told Bill is years old, intelligent but unimaginative, and compulsive. In school, he was strong in mathematics but weak in humanities.
Options included: "Bill plays jazz for a hobby" vs. "Bill is an accountant who plays jazz for a hobby."
Participants consistently ranked the conjunction as more likely because the description of being "unimaginative and mathematical" fits the stereotype of an accountant.
Case Study of Linda (The Feminist Bank Teller)
Linda's description: years old, single, outspoken, and very bright. As a student, she was deeply concerned with issues of discrimination and social justice.
Participants estimate her likelihood of being a bank teller (A) versus a bank teller who is active in the feminist movement (A + B).
Linda’s description creates a high representativeness for the "feminist" category, leading participants to ignore the logical nesting of categories.
Findings and Probability Correlation
Research Outcomes and Statistical Trends
Kahneman and Tversky found a near-perfect correlation between the ranking of representativeness (similarity) and the ranking of probability.
The correlation was reported at for the Bill study and for the Linda study.
These findings prove that humans often substitute a difficult question (What is the probability?) with an easier one (How similar is it?).
Conjunction Fallacy
Definition and Formal Logic
Conjunction Fallacy: The logical error of assuming that specific conditions are more probable than a single general one.
Logical Inequality: For any two events and , the probability of the conjunction () cannot exceed the probability of either individual event (). Formally: .
Explanation using Set Theory
The intersection of two sets (feminist bank tellers) is always a subset of the larger set (all bank tellers). Therefore, it is statistically impossible for Linda to be more likely to be a "feminist bank teller" than just a "bank teller."
Additional Research and Misinterpretations
University of British Columbia Findings
Despite being presented with the logic, of students at UBC failed the Linda problem.
Even when researchers explicitly separated the options to make the logical hierarchy transparent, the majority of subjects still chose the more "representative" option.
This suggests that representativeness is an "automatic" System 1 process that is difficult for the "logical" System 2 to override.
General Misunderstandings and Psychological Mechanisms
Insensitivity to Base Rates (Base Rate Neglect)
People often ignore the general prevalence (base rate) of a category in the population.
Example of Jack: Jack is a -year-old man who likes mathematical puzzles and has no interest in social issues. Even if he works in a firm with lawyers and only engineers, people will guess he is an engineer based on his traits, ignoring the odds that favor him being a lawyer.
Misconception of Chance (Law of Small Numbers)
People expect a sequence of events generated by a random process to represent that process even in short sequences.
Gambler's Fallacy: In coin flipping, people think "H-T-H-T-T-H" is more likely than "H-H-H-H-H-H" because the former looks more "random," even though both sequences have a probability of .
Ongoing Studies and Their Implications
Richard and Social Stereotypes
Modern research uses sketches of Richard (varying traits) to see if conjunction fallacies occur with negative stereotypes (e.g., "Richard is a criminal" vs. "Richard is a Christian criminal").
These studies aim to see how emotional valence affects probability estimation, showing that strong stereotypes (like a "rapist" or "teacher") skew logic even further than neutral roles.
Wrap Up and Theoretical Significance
Final Discussion Points
The prevalence of the conjunction fallacy and base rate neglect highlights the dual-process theory of cognition (System 1 vs. System 2).
Education on statistics does not always eliminate these biases; the intuitive "feel" of representativeness often triumphs over mathematical training.
Understanding these heuristics is critical for professions requiring risk assessment, such as medicine, law, and finance.